posted 13 November 2006 19:36            

I've been comparing Kyma to various soft samplers and I must say I'm amazed at how much better Kyma sounds next to some popular samplers in Logic. All in digital domain... My uneducated guess would be the soft samplers makes shortcuts to preserve CPU or something... I won't be able to use any of them again...

posted 14 November 2006 02:10            

There are multiple ways of controlling several Capys at once. The only thing is that you will always need one computer attached to one Capybara.
You can do this with the MIDI protocol, but my prefered way would be to use the OSC protocol over an ethernet network. For this, you can use the OSCulator, an applicatio I wrote to control the Capybara from OSC. It connects to the Flame interface over FireWire, and translates OSC
messages to control change messages that the Capybara can understand.

You can find it here :
http://web.mac.com/tuscland/iWeb/OSCulator/Welcome.html

Best from,
Camille

posted 14 November 2006 03:58            

What I had in mind was to have two or three Capys connected to one computer in order to get more I/Os and a good workflow of tweak/compile/upload/try-out. I was wondering if it's possible to connect more than one Flame to one computer and select which Capy to upload to, or if I'd have to run separate machines...

posted 14 November 2006 04:52            

No, you can't have one computer for multiple capys.
You must have one Capybara, for one computer.
If you really want to command multiple capybara with multiple computers, I would suggest you use Apple Remote Desktop or VNC in order to share the mutliple screens on the same principal screen.
Are you sure you need so much processing?
A fully loaded Kyma isn't it enough?
I am curious to know your needs!

posted 14 November 2006 14:13            

Then I would also like to have 24 I/Os to run separate inputs live into a DAW.

I just realized how good Kyma sounds even just playing back samples so I want to move as much as possible of the sample playback to the Capy... I've been using soft samplers in Logic and it is very convenient - little to keep track of in terms of recall etc... I'm learning how to incorporate Kyma in this way - keep track of changes and be able make changes quickly.

But anyway... a roundabout way to explain what I'm trying to get to multiple Capys from the same computer with a means to switch between them would have been fab. Second best is probably a Mac mini tied to each, networked and then controlled with Remote Desktop. Or maybe there's a keyb/mouse/monitor switcher where I can go between three computers...?

posted 14 November 2006 18:01            

I'm also wondering (a lot of wondering here...) how to get more predictable timing playing the Capy from Logic. "Plug 1 (Capybara I/O}" shows up as an available MIDI out but the timing jitter is > 100 samp on my system which is just too much. Straight MIDI is a bit better: +- 30 samp or so... all from a quick and dirty test. Still not workable... I still haven't successfully set up and used the Flame MIDI. The way my system is setup I can't have the Capy be the master clock and I read somewhere that is a prerequisite.

ATB,
Olle

posted 15 November 2006 05:38            

If your question is to find a way to use OSCulator with a sequencer, the answer is that you should use an OSC enabled sequencer. I think the fewer steps there are between the emission and the reception of the control signal, the less latency you will get.

Frankly, I am not sure you will have better results when using through OSCulator. I tried to design OSCulator in order to have shortest and most efficient signal pathes, but there is still a small jitter added at every stages, and I have no control over it : OSC message sender -> IP network transmisson -> OSCulator -> FireWire -> Capybara.
It would be very interesting to do some timing tests to make sure the latency is not so problematic. If the results are good, then I think I can spend some time to write code in OSCulator that will handle note-on/note-off messages.

A bunch of questions for you:
I would like to know when you are talking about "jitter" do say "clock stability" or "latency"?
Could you please describe how you did test the jitter of the MIDI interface (physical and Capybara I/O)? Don't forget that in Kyma there is a 3ms latency (that you can control in the Preferences pane).

I will make some tests on my side and see wether our results match.
A latency of 30 samples is very good for a physical MIDI interface, remember that the MIDI standard is very old.

If you may please describe a little bit more the application you are trying to achieve, then I could perhaps find another solution for you.

posted 16 November 2006 19:05            

quote:

---

Originally posted by armand:
I wonder what's the story behind this Capy family? http://1106.suac.net/SS2001/photo/DSC00245.JPG

---

posted 17 November 2006 02:34            

After a discussion with Kurt, I learned that the MIDI to FireWire driver (coded by Apple) that runs the "Capybara I/O Plug 1" MIDI port is rate limited to 3 kHz, in order to comply with the standard (or oldest MIDI interfaces).

I found out that there are two more ways of sending MIDI informations to Kyma with OSCulator, but I need to do some coding.
1. In a DAW, with a AudioUnit Instrument which translate the MIDI message to OSC, send translated informations to OSCulator, then in OSCulator forward them to the Capybara. Or use a MIDI to OSC translator to forward the messages to OSCulator.
2. Directly integrate a MIDI input in OSCulator and send full rate note-on/off events and control changes to Kyma.

I'd like to have some feedback about this and know if some users have preferences over the two methods.

posted 17 November 2006 05:38            

Unfortunately not, for the moment OSCulator is still one way, and will become two way when the Flame firmware will allow bi-directional communications.

I am talking about the direction that goes to Kyma.

posted 19 November 2006 16:37            

Camille, thanx for taking the time out to discuss this. Much appreciated

posted 20 November 2006 03:10            

Now I clearly understand what you mean by jitter. I guess you need a robust timing in order to be able to shift in time the processed audio by the same amount.
Could you please tell me what is a "acceptable" jitter for you ?

Would you accept to test a AudioUnit instrument that converts incoming MIDI on a track to Kyma commands? I think this is the most straightforward way to get the proper timing. If yes, please send me you private email, mine is tuscland@mac.com.

I like the idea of SMPTE sync, and it is something to study, but I fear this will be difficult to implement cleanly, no guarantee, thus not predictable ...

Recently, I have done big improvements in OSCulator, but I can't garantee a better jitter than about 500 samples right now. OSCulator is good for controlling, not for timing (yet).

posted 20 November 2006 03:57            

Acceptable 'timing jitter' would probably be < 20 samp. Still not quite there for percussive stuff but other types of sounds would work well. Is that theoretically possible?

I remember reading a few years ago MIDI is treated with low priority in OSX. Do you know if that's still the case?

posted 20 November 2006 04:21            

Cool, I will build this AudioUnit, just allow me a fews days to be sure I send you something that works nicely.

About MIDI, Kurt told me that the MIDI driver used for "Plug 1 (Capybara I/O)" was a bulk driver written by Apple. They decided to rate limit the MIDI to 3000 events/sec in order to comply with the older standard.

However, I believe MIDI is not rate limited internally, I don't see any reason why they would do this.

posted 20 November 2006 12:25            

quote:

---

Originally posted by armand:
Well if you look at the TurboMIDI interface from Elektron it is possible to speed up MIDI, and it also works with standard MIDI support. http://www.elektron.se/stories/?y=2006&p=3

---

Would the MIDI interface in the Capy be able to keep up with this?

posted 24 November 2006 03:25            

Did you receive my email?